Recommendations and Approaches to Marketing Analytics

Have you ever heard of six degrees of separation if not let this fun Ted Talk explain this in a fun video:

Caution this video does contain some sensitive language.

Social network analysis identifies relationships, influencers, information dissemination patterns, and behaviors among connections in a network.  Social network analysis results in visual maps that trace connections in the population and ultimately represent the size and structure of the networks.

Social networks provide an opportunity for companies to communicate with, and start conversations with, customers.  Social network analysis ultimately provides a picture of the network.  Companies can use this picture to better understand communities, influencers, and conversations that emerge.  Social media interactions among participants in the network develop organically from the company’s original post.  Other companies also maintain brand pages.  Some companies recruit influencers from their followers, like Sephora, while others collaborate or partner with already entrenched influencers.  Regardless of the strategy, interacting with existing and potential customers through networks can introduce products or services, increase brand awareness, and improve sales.  Using social network analysis, companies can monitor conversations about brands and relationships occurring from those interactions.  Using these insights, companies can better understand consumption behavior and brand preferences.

Measures of centrality indicate the influence a node has in the network and also a node’s strategic network position.  Degree centrality measures centrality based on the number of edges that are connected to the node.  If the network is directed, there are two measures of degree: indegree and outdegree.  Indegree is the number of connections that point in toward a node.  Outdegree is the number of arrows that begin with the node and point toward other nodes.  Nodes with a higher degree of centrality have more links and are more central.  Betweenness centrality measures centrality based on the number of times a node is on the shortest path between other nodes.  Betweenness assesses positional centrality, and it shows which nodes serve as bridges between nodes in the network.  This measure helps identify individuals who influence the flow of information in the social network.  Eigenvector centrality measures the number of links from a node and the number of connections those nodes have.  It shows whether a node is well-connected to other nodes, who in turn are also well-connected.  This is a useful measure to identify individuals with influence over the network, not just the individuals directly connected to them.  The higher the eigenvector centrality value assigned to the node, the more the node has influence over the entire network.

Marketers may want to predict the next most likely link to be established in the network and link prediction may help.  With link prediction, the objective is to predict new links between unconnected nodes.  Link prediction uses a variety of methods such as similarity and machine learning algorithms.  When nodes are closer together, the more likely there will be a relationship between them.  Using link prediction, future associations that are likely to occur can be more accurately predicted.